Intermediate Forms - Different compilers use different IFs
- Many compilers use more than one IF internally, though in the common two-pass organization one of these is distinguished as “the” intermediate form
- connection between the front end and the back end.
- the syntax trees passed from semantic analysis to intermediate code generation constitute a high level IF
- control flow graphs containing pseudo-assembly language (passed in and out of machine-independent code improvement) are a medium level IF
- the assembly language of the target machine serves as a low level IF
- Compilers that have back ends for different target architectures do as much work as possible on a high or medium level IF
- the machine-independent parts of the code improver can be shared by different back ends
Intermediate Forms - The back end of Figure 15.1 is too complex to present in any detail in a single chapter
- To limit the scope of our discussion, we will content ourselves in this chapter with producing correct but naive code
- This choice will allow us to consider a significantly simpler back end.
- Starting with Figure 15.1, we drop the machine-independent code improver and then merge intermediate and target code generation into a single phase
- generates linear assembly language - no code improvements for control flow, therefore, there is no need to represent that flow explicitly in a control flow graph
Code Generation - We also adopt a much simpler register allocation algorithm
- operates directly on the syntax tree prior to code generation - eliminates need for virtual registers and the subsequent mapping onto architectural registers
- Finally, we drop instruction scheduling. The resulting compiler structure appears in Figure 15.5.
- Its code generation phase closely resembles the intermediate code generation of Figure 15.1.
- An Attribute Grammar for GCD Example is presented in Section 15.3.1
Do'stlaringiz bilan baham: |
